OpenAI Launches GPT‑5.6‑Cyber as Daybreak Expands Advanced Defensive AI Access
OpenAI introduced GPT‑5.6‑Cyber for vetted defensive security researchers through Daybreak Red, targeting advanced vulnerability research and zero-day discovery.
Why it matters: Specialized frontier AI for defensive cybersecurity could materially change enterprise vulnerability research while raising dual-use governance questions.
OpenAI has introduced GPT‑5.6‑Cyber, a specialized artificial-intelligence model designed for advanced cybersecurity research and available to approved users through the company's restricted Daybreak Red program. The model is built for vulnerability discovery, exploit validation and other advanced defensive workflows. OpenAI says researchers have used the model to identify previously unknown vulnerabilities, including flaws in Google's V8 JavaScript engine. The company rates the model at its High cybersecurity-capability threshold but below Critical and restricts access to vetted users.